done

README.md

@@ -28,101 +28,98 @@ configs:
     path: card_samples/parquet/qa.parquet
 ---
 
-# KnowCP
+# KnowCP Dataset Repository
 
-KnowCP is a benchmark for Chinese painting understanding and reasoning.
-It supports recognition-style tasks (region/text extraction) and knowledge/reasoning tasks (open QA, multiple choice, and multi-turn QA).
+Project homepage and benchmark details:
+https://41-edu.github.io/KnowCP-Benchmark/
 
-## Dataset Summary
+Script repository:
+https://github.com/41-edu/KnowCP
 
-- Name: KnowCP
-- Modality: image + text
-- Language: Chinese
-- License: CC BY 4.0
-- Total images: 2331
-- Total question items (full set): 26137
+## Project Overview
 
-## Dataset Viewer Layout
+KnowCP is a comprehensive benchmark for evaluating multimodal large language models on Chinese painting understanding, from basic recognition to deep reasoning.
 
-The dataset card has two simple series:
+This repository is the dataset repository of KnowCP. It stores the image corpus, annotation resources, question sets, and knowledge-base metadata used by the benchmark.
 
-- `image`: first 100 single-image artworks with metadata
-- `qa`: 14 question types, 10 items per type (140 total)
+## Dataset Scale
 
-The card data is generated only from core folders:
+The following counts are aligned with the benchmark website configuration and public content files.
 
-- `images/`
-- `kb/`
-- `annotations/`
-- `questions/`
+1. Paintings: 1210
+2. Images: 2331
+3. Seal annotations: 8690
+4. Colophon or inscription annotations: 2434
+5. Object or element annotations: 5052
+6. Technique annotations: 1312
+7. Total question items used by the benchmark website: 38044
 
-## Repository Structure
+## Repository Contents
 
-- `images/**`: image assets
-- `kb/knowledge_base.json`: artwork metadata (`title`, `artist`, `dynasty`, `institution`)
-- `annotations/*.json`: seal/inscription/object/technique annotation exports
-- `questions/by_type/*.jsonl`: source questions by type
-- `card_samples/parquet/image.parquet`: image series table
-- `card_samples/parquet/qa.parquet`: qa series table
+1. images
+  All painting images and related sub-images used by the benchmark.
 
-## Data Fields in Viewer
+2. questions
+  All benchmark QA files. We provide 14 question files under [hf_repo/questions/by_type](hf_repo/questions/by_type), and each file corresponds to one question source type.
 
-### `image`
+  All QA content is in Chinese.
+  For concrete QA presentation style and English-facing examples, see:
+  http://localhost:5173/#question-distribution
 
-- `image` (rendered image)
-- `image_id`
-- `title`
-- `artist`
-- `dynasty`
-- `institution`
-- `material`
+  14 question files and counts used in the website benchmark view:
 
-### `qa`
+  - ITT_MHQA_choice.jsonl: 4232
+  - ITT_MHQA_fillin.jsonl: 4232
+  - MITT_MHQA_choice.jsonl: 608
+  - MITT_MHQA_fillin.jsonl: 608
+  - TTI.jsonl: 1210
+  - SR.jsonl: 1792
+  - IR.jsonl: 2351
+  - ER_choice.jsonl: 11359
+  - ER_fillin.jsonl: 5052
+  - TR_choice.jsonl: 1312
+  - TR_fillin.jsonl: 1312
+  - VA.jsonl: 922
+  - CC.jsonl: 922
+  - PR.jsonl: 922
 
-- `image` (rendered image)
-- `qid`
-- `type`
-- `question`
-- `ground_truth`
+  Simple task descriptions, following the website organization:
 
-## Usage
+  Foundational Knowledge
+  - ITT and MITT style tasks: identify title-level information from one image or multiple images.
+  - TTI: retrieve the matching image from title information.
+  - MHQA: multi-step reasoning over image and context, provided in choice and fill-in formats.
 
-```python
-from datasets import load_dataset
+  Visual Content
+  - SR: seal recognition.
+  - IR: inscription or colophon recognition.
+  - ER: element recognition, provided in multiple-choice and fill-in formats.
+  - TR: painting technique recognition, provided in multiple-choice and fill-in formats.
 
-repo_id = "g41/KnowCP"
+  Deep Reasoning
+  - VA: visual analysis.
+  - CC: cultural context reasoning.
+  - PR: provenance research reasoning.
 
-image_ds = load_dataset(repo_id, name="image")
-qa_ds = load_dataset(repo_id, name="qa")
+3. annotations
+  Fine-grained annotations produced by our annotators for each painting image, including seals, inscriptions, elements, and techniques.
 
-print(image_ds["train"][0]["title"])
-print(qa_ds["train"][0]["qid"])
-```
+  Annotation visualization references:
+  http://localhost:5173/#distribution
 
-To regenerate the dataset-card samples:
+4. kb
+  Core metadata per painting, including identity and background attributes used by benchmark tasks.
 
-```bash
-python for_hf/scripts/build_dataset_card_samples.py \
-  --hf-repo hf_repo \
-  --out-dir hf_repo/card_samples \
-  --images-limit 100 \
-  --qa-per-type 10
-```
+## Folder Reference
 
-## Evaluation with This Project
+- [hf_repo/images](hf_repo/images): all images
+- [hf_repo/questions](hf_repo/questions): question sets
+- [hf_repo/annotations](hf_repo/annotations): fine-grained annotations
+- [hf_repo/kb](hf_repo/kb): painting metadata knowledge base
+- [hf_repo/mappings](hf_repo/mappings): mapping resources used in processing and alignment
 
-To run the full benchmark with local files:
+## If You Want to Run Evaluation
 
-```bash
-python github_repo/run_eval_for_hf.py \
-	--workspace-root D:/mylib/benchmark \
-	--items-dir D:/mylib/benchmark/hf_repo/questions/by_type \
-	--model bytedance-seed/seed-2.0-lite \
-	--api-base https://openrouter.ai/api/v1 \
-	--api-key YOUR_API_KEY
-```
-
-## Citation
-
-- Code: https://github.com/41-edu/KnowCP
-- Dataset: https://huggingface.co/datasets/g41/KnowCP
+This repository provides data only.
+For actual model benchmarking and score computation, please use the script repository:
+https://github.com/41-edu/KnowCP